Abstract
China Railway Corporation (CRC) has been paid more attention to passenger transportation revenue, with its increase proportion in transportation revenue. Due to the price regulation, the only way CRC can improve ticket sale profit is to find a best seat allocation scheme. This study focuses on the optimization of railway revenue management problem in China with consideration of i) customer behaviors including their arrival and purchase preferences, ii) a specific ticket booking mechanism called "seat-based control". To evaluate the performance of seat-based control, we build a Discrete-Time Markov Chain model to describe the ticket reservation process and then design a genetic algorithm to find approximate solutions. The performance of proposed method is tested in two experiments with two other benchmarks. Finally, we apply it to practical data of the Nanning-Guangzhou high-speed railway line.